Validation of Surface Irradiance Calculated in NASA's Clouds and the Earths Radiant Energy System (CERES) SYN and EBAF Data Products

NASA's Clouds and the Earths Radiant Energy System (CERES) project archives several data products that provide global, gridded and temporally complete maps of Top of Atmosphere (TOA) and surface radiative fluxes. The Ed3 SYN1Degree product releases 3 hourly, daily, and monthly mean estimates of fluxes at 5 levels (TOA, surface, 500 hPa, 200 hPa, and 70 hPa) in the atmosphere. The Energy Balanced and Filled (EBAF) product has monthly temporal and 1°×1° spatial resolutions and is restricted to Top Of Atmosphere (TOA) and surface fluxes. The advantage of the Ed2.7 EBAF product being that it uses a constrainment algorithm to balance the net TOA fluxes against estimates of the energy absorbed into the global oceans (July 2005 through 2010 mean of 0.58 Wm-2). The monthly mean surface EBAF fluxes, which are initially derived from the SYN data product, are consistent with TOA EBAF. This presentation provides an evaluation of the SYN and EBAF surface fluxes by comparing them with observed shortwave (SW) and longwave (LW) downwelling fluxes at 85 sites across 13 years. We use data from 37 land sites and 47 ocean buoy locations from Mar 2000 through Dec 2012. Note that not all surface sites include observations across entire time span. Evaluation results are presented for 3 hour, daily, and monthly mean temporal resolutions and include other comparable radiative fluxes from re-analyses such as MERRA and ISCCP-FD.

Specifically, this work presents direct comparisons of observed and calculated surface downwelling LW and SW fluxes analyzing anomaly time series, mean statistics and diurnal cycle. Surface fluxes computed with and without the inclusion of Geostationary satellite (GEO) derived fluxes and clouds indicate modeled diurnal cycle results improve with the addition of the GEO data. Monthly mean comparison results include for mean downward LW flux of 330Wm-2, SYN has a bias of -4.1Wm-2 and a monthly standard deviation of the bias 10.6Wm-2. The EBAF surface LW calculations resulted in a mean bias of 0.4Wm-2 with a standard deviation of 10.2 Wm-2. For surface downwelling SW the mean observed flux was 201Wm-2 with the SYN and EBAF having bias of 2.2Wm-2 and -0.1Wm-2 respectively. Their standard deviations for the SW down at the surface are 11.8Wm-2 for both data products. Overall we find that the SYN and EBAF results provide better comparisons to surface observed fluxes than the other re-analyses products. This is attributed to the fact that the multiple geostationary data sets are calibrated to the CERES Terra and Aqua instruments across the years providing TOA fluxes that are consistent with the cloud properties derived from the narrowband instruments. In addition, TOA flux constraint in computing surface fluxes reduces the bias of downwelling SW surface fluxes.